Decentralized nonlinear model predictive control of multiple flying robots

In this paper, we present a nonlinear model predictive control (NMPC) for multiple autonomous helicopters in a complex environment. The NMPC provides a framework to solve optimal discrete control problems for a nonlinear system under state constraints and input saturation. Our approach combines stabilization of vehicle dynamics and decentralized trajectory generation, by including a potential function that reflects the state information of possibly moving obstacles or other vehicles to the cost function. We present various realistic scenarios which show that the integrated approach outperforms a hierarchical structure composed of a separate controller and a path planner based on the potential function method. The proposed approach is combined with an efficient numerical algorithm, which enables the real-time nonlinear model predictive control of multiple autonomous helicopters.